Opioid agonists acting at their receptors alter intracellular events by initiating activation of various types of Gi/Go proteins. This can be measured by the binding of the stable GTP analog [35S]guanosine-5′-O-(3-thio)triphosphate ([35S]GTPγS). In this study agonist efficacy is defined by the degree to which an opioid stimulates the binding of [35S]GTPγS. This allows for a definition of full and partial agonists; a full agonist causing a greater stimulation of [35S]GTPγS binding than a partial agonist. The hypothesis that the rate of agonist-stimulated [35S]GTPγS binding is dependent upon agonist efficacy was tested using membranes from C6 glioma cells expressing μ- or δ-opioid receptors. At maximal concentrations the rate of agonist-stimulated [35S]GTPγS binding followed the efficacy of μ-agonists in stimulating [35S]GTPγS binding, i.e., [d-Ala2,N-Me-Phe4,Gly5-ol]-enkephalin > morphine > meperidine > butorphanol > nalbuphine. At submaximal concentrations of μ- or δ-full agonists the [35S]GTPγS association rate was also reduced, such that the rate of [35S]GTPγS binding correlated with the extent of [35S]GTPγS bound, whether this binding was stimulated by a full agonist or a partial agonist. Agonists also stimulated [35S]GTPγS dissociation, showing that binding of this stable nucleotide was reversible. Comparison of the δ-agonists [d-Ser2,Leu5]-enkephalin-Thr and (±)-4-((α-R*)-α-((2S*,5R*)-4-allyl-2,5-dimethyl-1-piperazinyl)-3-hydroxylbenzyl)-N,N-diethylbenzamide, a compound with slow dissociation kinetics, showed the measured rate of G protein activation was not influenced by the agonist switching between receptors. The results are consistent with the idea that the active state(s) of the receptor induced by full or partial agonists is the same, but the number of activated receptors determines the rate of G protein activation.
Opioid agonists alter intracellular metabolism by binding to membrane-bound receptors and initiating activation of various types of Gi/Go proteins. The rate-limiting step in this activation is the dissociation of the GDP that is bound to the Gα subunit of the G protein under resting conditions. This allows GTP to bind and the active Gα-GTP and βγ to dissociate and interact with downstream effectors. The activity of agonists acting on such systems can be determined by measuring the agonist-stimulated binding of the stable GTP analog [35S]GTPγS, to α-subunit proteins (Traynor and Nahorski, 1995). The potency of an agonist in this system is then defined by its EC50 for stimulating [35S]GTPγS binding and the efficacy of an agonist by the degree to which it maximally stimulates [35S]GTPγS binding. Thus, in the context of this article, efficacy is defined as the relative ability of an opioid agonist to stimulate [35S]GTPγS binding, with full agonists stimulating [35S]GTPγS to a maximal level and partial agonists causing a reduced level of [35S]GTPγS binding. The relative efficacy of a partial agonist is then expressed as a fraction of the full agonist response.
The differential ability of agonists to activate heterotrimeric G protein-linked receptors has been attributed to the affinity of ligand bound-receptor for G protein (Tota and Schimerlik, 1990) as well as to the ability of the agonist bound-receptor to initiate GDP dissociation from G protein (Breivogel et al., 1998). Although the relative orientation of receptors and G proteins and the surfaces through which they interact with one another have been partially defined (Bourne, 1997) our understanding of protein conformational changes that may account for differences between full and partial agonists is limited (Burgen, 1981; Kenakin, 1995a,b).
Membranes prepared from C6 glioma cells stably expressing the rat μ- and δ-opioid receptors are an excellent model system to evaluate the relative ability of opioid agonists to stimulate [35S]GTPγS binding (Lee et al., 1999). In membranes from these cells opioid partial agonists stimulate [35S]GTPγS binding at a considerably slower rate than full agonists, when both are at concentrations causing maximal effects (Remmers et al., 2000). This lower rate of G protein activation by partial agonists can be explained either by a lower number of receptors in the same active conformational state, or by the fact that different conformational states are formed in the presence of partial agonists. A partial agonist-specific state of the receptor would have a lower affinity for G protein, and/or cause a slower dissociation of GDP.
In this study the hypothesis that full agonists and partial agonists induce the same active state (or states) of the receptor, but that the fraction of receptors in these active states is less with partial agonists has been examined. The rates at which opioid full and partial agonist stimulate the binding of [35S]GTPγS to membranes from C6μ cells have been determined. The results show that the association rate is related to efficacy only at maximal receptor occupancy and that rate is more correctly correlated with the degree of stimulation of [35S]GTPγS binding, whether this is caused by a full or partial agonist. These findings are interpreted to suggest that partial agonists stabilize the same active conformation state(s) of the receptor as full agonists, but that the number of receptors in these active states is much reduced.
[35S]GTPγS (1300 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA). DAMGO and DSLET were purchased from Sigma Chemical (St. Louis, MO), and nalbuphine, naltrexone, morphine, meperidine, and butorphanol were obtained through the Opioid Basic Research Center at the University of Michigan (Ann Arbor, MI). Dulbecco's modified Eagle's medium, fetal bovine serum, and Geneticin were from Invitrogen (Carlsbad, CA). Trizma base, GDP, GTPγS, and all other biochemicals were purchased from Sigma Chemical.
C6 glioma cells stably transfected with the μ- or δ-opioid receptor (Lee et al., 1999) were grown to confluence under 5% CO2 in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and either with 0.5 mg/ml Geneticin (for subculture) or without Geneticin (for harvest). The cells were typically subcultured at a ratio of 1:20 to 1:30 with partial replacement of the media on day 4 and the day before subculturing or harvesting at day 5 or 6 for C6δ cells and day 6 or 7 for C6μ cells.
Cells were washed two times with ice-cold phosphate-buffered saline (0.9% NaCl, 0.61 mM Na2HPO4, pH 7.4) then detached from flasks by incubation in lifting buffer (5.6 mM glucose, 5 mM KCl, 5 mM HEPES, 137 mM NaCl, 1 mM EGTA, pH 7.4) at room temperature and pelleted by centrifugation at 200g for 3 min. The cell pellet was resuspended in 10 volumes of ice-cold 0.32 M sucrose (pH 7.4 with 1 mM Tris-HCl) with a Teflon-glass Dounce mounted to a Tri-R Stir-R motor at 1000 rpm. The suspension was then centrifuged for 10 min at 1000g at 4°C, and the supernatant was removed and kept on ice. The resuspension and centrifugation were repeated with the remaining pellet an additional three times, saving the supernatant from each spin in tubes kept on ice, to further break up the cells and increase the yield. The combined supernatants were then centrifuged at 15,000g for 20 min at 4°C. The resulting supernatant was discarded and the upper pellet was separated from the lower pellet by gently washing with ice-cold 0.32 M sucrose. The upper pellet suspension was homogenized with a glass-glass Dounce and centrifuged at 15,000g for 20 min at 4°C. The upper pellet was resuspended in 50 mM Tris-HCl buffer, pH 7.4, and centrifuged 20 min at 20,000g and 4°C. The final pellet was resuspended in 50 mM Tris buffer and frozen at −80°C in 0.25-ml aliquots (1–2 mg/ml).
Protein concentration was determined by the method of Bradford (1976) with a bovine serum albumin standard. Samples were dissolved with 1 N NaOH for 30 min at room temperature and neutralized with 1 M acetic acid before protein determination.
[35S]GTPγS Association Rate.
[35S]GTPγS binding was determined in the presence of 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM dithiothreitol (added fresh), 50 μM GDP, 80 to 960 μg of membrane protein (5–60 μg/0.5-ml sample so less than 10% of the added [35S]GTPγS is bound), an appropriate concentration of agonist ligand or double distilled H2O, and 0.05 nM [35S]GTPγS in a total volume of 8.0 ml. Membranes and assay buffer were preincubated in the absence of ligand and [35S]GTPγS for 10 min in a shaking water bath at 37°C. Ligand or ddH2O was added and the tubes were incubated for an additional 10 min (or 25 min with lower drug concentrations) at 37°C. [35S]GTPγS was added to initiate binding. The association was terminated by removing 0.5-ml samples from each tube at 2- to 10-min intervals up to 100 min and by filtering through glass fiber filters (Schleicher & Schuell 32; Schleicher & Schuell, Keene, NH). The filters were quickly rinsed four times with 2 ml of ice-cold 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, and 5 mM MgCl2. Filters were placed in polypropylene vials with 0.4 ml of ethanol and 4 ml of Ultima Gold scintillation cocktail was added and the samples subjected to liquid scintillation counting. Agonist-stimulated [35S]GTPγS binding was determined as the difference between [35S]GTPγS binding with and without ligand at each time point. The data were fit to a one-phase exponential association curve (y = ymax (1 −e −k · x) wherey is binding increasing to a maximum plateau (ymax),x is time, and k is the rate constant, using GraphPad Prism (version 3.0; GraphPad, San Diego, CA) to determine the rate and maximum agonist stimulated [35S]GTPγS binding. Each experiment was repeated two to three times.
Agonist-Stimulated [35S]GTPγS Dissociation Rate.
[35S]GTPγS was bound in the presence of 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM dithiothreitol (added fresh), 50 μM GDP, 200 to 300 μg of membrane protein (10–15 μg/0.1-ml sample), ligand (maximally efficacious concentration), and 0.08 nM [35S]GTPγS in a total volume of 2.0 ml. Membranes and assay buffer were preincubated in the absence of ligand and [35S]GTPγS for 5 min in a shaking water bath at 25°C. Ligand or ddH2O was added and the tubes were incubated for an additional 10 min at 25°C, followed by the addition of [35S]GTPγS. Binding was allowed to proceed for 80 min at 25°C, followed by the addition of 10 μM naltrexone or ddH2O. After 5 min, 50 μM GTPγS was added to initiate dissociation and 0.1-ml samples were removed from each tube at 1- to 15-min intervals up to 58 min and processed as described above. The data were best fit to a one-phase exponential decay curve by using the formula in GraphPad Prism (version 3.0), i.e., y = span · e −k · x + plateau, where x is time, y is binding that starts out as equal to span + plateau and decreases to plateau with a rate constant k. Each dissociation experiment was repeated five times.
The rate of G protein activation by maximally effective concentrations of agonists was determined. GDP (20 min) and agonist (10 min) were preincubated with membranes at 37°C, so that the rate observed was not influenced by the rate of agonist or GDP binding. In addition, ligand depletion was kept below 10% by using large incubation volumes. Agonist-stimulated [35S]GTPγS binding in C6μ membranes increased in a time-dependent manner and reached a plateau (Fig.1). The association kinetics best fit to a one-component exponential association. The extent of maximal agonist-stimulated [35S]GTPγS binding was smaller for nalbuphine (partial agonist) than for the DAMGO (full agonist). When maximal concentrations of each ligand were used the rate of [35S]GTPγS binding was also slower for the nalbuphine. Slower association rates were also obtained using maximally effective concentrations of the μ-partial agonists morphine, meperidine, butorphanol, and nalbuphine (Table1).
The rate of [35S]GTPγS binding reflects the rate of GDP dissociation from the G protein guanine nucleotide binding site. In addition, prebound [35S]GTPγS can also be dissociated from G protein by agonist action (Hilf et al., 1992). Thus, the activated receptor can function as a guanine nucleotide exchange factor. In this case one would predict that the rate of 35S-labeled guanine nucleotide dissociation would be faster in the presence of the full agonist DAMGO compared with the partial agonist nalbuphine. Both DAMGO and nalbuphine increased the rate of [35S]GTPγS dissociation in membranes compared with the dissociation rate seen when agonist action was blocked by the antagonist naltrexone. The half-time for the DAMGO-mediated dissociation was 2.5 min compared with 6.4 min for that of nalbuphine (Fig. 2).
To relate the extent of receptor activation to the observed rate of agonist-stimulated guanine nucleotide binding, the binding of [35S]GTPγS was studied using submaximal concentrations of the agonist DAMGO. Both 20 and 50 nM DAMGO stimulated [35S]GTPγS binding but displayed both a diminished rate and extent of [35S]GTPγS binding compared with a maximally effective concentration of DAMGO (Table 2). The length of the preincubation time with agonist was increased (to 25 min at 37°C) so that the rate observed was not indicative of the rate of agonist binding to the receptor with these lower concentrations of ligand. There was a strong correlation between the degree of [35S]GTPγS stimulation, afforded by either partial or full agonists, and the rate of G protein activation (Fig.3).
It is possible that the slow rate of G protein activation with partial agonists and submaximal concentrations of full agonists is caused by the need for the agonist to switch from one receptor to another to activate G proteins available to several μ-receptors. To test this hypothesis, an efficacious agonist that displays very high affinity for the receptor even in the presence of sodium and guanine nucleotide was required. BW373U86 is a nonpeptide δ-opioid receptor agonist that is insensitive to sodium and GDP (Childers et al., 1993; Wild et al., 1993). In membranes prepared from C6δ cells, BW373U86 binding affinity for the receptor was 0.45 nM under conditions of the [35S]GTPγS assay (50 μM GDP, 150 mM NaCl;Remmers et al., 1999). Assuming a bimolecular association rate of 1 × 106 M−1s−1, this binding affinity corresponds to a half-time for dissociation of 26 min. Thus, this ligand will remain associated with the receptor during a majority of the data acquisition time. In contrast, the dissociation rate for the δ-peptide agonist DSLET is fast and is highly sensitive to the presence of sodium ions and guanine nucleotides (Childers et al., 1993). The apparent rate of [35S]GTPγS binding was not significantly different whether maximally stimulating concentrations of the full agonists BW373U86 or DSLET were present in the assay (Table 2). In addition, using half-maximal agonist concentrations, the rates of BW373U86- and DSLET-stimulated [35S]GTPγS binding were not significantly different from each other, although the rates were slower than observed with maximal agonist concentrations (Table 2).
The rate-limiting step in the activation of G protein after agonist occupation of a seven transmembrane domain receptor is the dissociation of GDP, yielding the ternary complex of ligand-receptor-G protein where G protein is unliganded (Gempty). Subsequent GTP binding to Gempty occurs very fast (Ferguson et al., 1986;Krumins and Barber, 1997). Thus, the ability of agonist to stimulate GDP dissociation from the G protein can be determined by measuring the rate of GTP association, or the association of the labeled analog [35S]GTPγS. In the present experiments agonist was preincubated with receptor so that agonist binding did not influence the measured [35S]GTPγS association rate.
μ-Opioid partial agonists, that is, compounds that displayed a decreased extent of G protein activation at full receptor occupancy compared with the full agonist DAMGO, also showed a reduced rate of [35S]GTPγS binding. These findings are in agreement with the idea that different drugs cause different active conformations of the receptor, leading to differences in the rate of GDP release. This is consistent with results in the cannabinoid system that the mechanism of agonist efficacy in stimulating [35S]GTPγS binding is the magnitude of the decrease in the affinity of G protein for GDP (Breivogel et al., 1998). The findings could also be explained by a decreased affinity of the partial agonist-bound receptor for G protein compared with full agonist-bound receptor; such as is observed with reconstituted M2 receptor coupling to Gi (Tota and Schimerlik, 1990). In support of these findings there are several examples in the literature suggesting ligand-specific conformational states of agonist-occupied seven transmembrane receptors (Gether et al., 1995; Büküsoglu and Jenness, 1996; Krumins and Barber, 1997; Berg et al., 1998).
Thus, it can be hypothesized that opioid partial agonists induce different active conformational states of the receptor than those induced by full agonists, and this is the reason for their lower efficacy. To test this hypothesis the number of receptors activated by DAMGO was manipulated by altering the concentration of this full agonist. Under these conditions comparison of the rates of [35S]GTPγS binding at the μ-receptor with the extent of [35S]GTPγS bound showed an excellent correlation regardless of whether the full agonist or a partial agonist was used. It is unlikely that agonist-occupied receptor conformations will be different depending upon the available concentration of the full agonist DAMGO and so this finding suggests that the extent of G protein activated, not the fact the some compounds are of lower efficacy, determines the rate of G protein activation. These data are consistent with the idea that the mechanism of GDP dissociation is the same for both full and partial agonists and therefore that the active conformation of the opioid receptor is the same regardless of whether the receptor is occupied by a full or partial agonist. A similar conclusion can be inferred from the ability of β-adrenergic agonists of differing efficacy to activate receptor phosphorylation (Benovic et al., 1988).
In C6μ cells we have previously shown the ratio of μ-receptors to G protein that can be maximally activated by these receptors is 1:4 (Remmers et al., 2000). These are all pertussis toxin-sensitive Go/Gi proteins. It is generally considered that [35S]GTPγS binding is irreversible due to the stability of the molecule to the intrinsic GTPase of the Gα subunits (Sternweis and Robishaw, 1984). If this is the case an important question is why partial agonists, and indeed full agonists at submaximal concentrations, cause a plateau effect and do not, given sufficient time, activate all available Go/Gi proteins. However, the dissociation experiments show that bound [35S]GTPγS is reversible in the presence of excess unlabeled GTPγS and that it can be driven off more rapidly by the full agonist DAMGO, in agreement with findings in the cannabinoid (Breivogel et al., 1998) and muscarinic systems (Hilf et al., 1992). Thus, the plateau represents a steady-state competition between [35S]GTPγS and GDP. Because a similar plateau can be reached by partial agonists and low concentrations of full agonist this suggests that they have similar effects on GDP and/or [35S]GTPγS binding and release. In addition, access of each activated receptor to all Go/i proteins in the cell is likely limited by cytoskeletal barriers or some other form of compartmentalization of receptors and G proteins (Neubig, 1998) as suggested by our previous studies in C6μ cells (Remmers et al., 2000). The fact that agonists stimulate [35S]GTPγS dissociation from the Gα subunits does suggest that the [35S]GTPγS bound G proteins remain accessible to the receptor.
The rate of turnover of agonist occupancy, with agonist moving between receptors, has been suggested to influence the rate of G protein activation by receptors (Stickle and Barber, 1996). However, the apparent rate of [35S]GTPγS binding was not significantly different with either BW373U86, a highly efficacious δ-opioid agonist that remains associated with the receptor during a majority of the data collection time, and the efficacious peptide DSLET that has more rapid receptor kinetics. This confirms that the ability of an agonist to activate many different receptors is not a significant determinant in the rate of G protein activation in the present experiments. Furthermore, the results establish that receptor kinetics is not a factor in the current analysis.
Consideration of the simple conformational selection model of receptor activation (Burgen, 1981) allows an explanation of the results obtained with the full and partial agonists. An agonist, by binding preferentially to an activated high-affinity state of the receptor (R*) will shift the equilibrium in favor of R* and so increase G protein activation. DAMGO has a 1200-fold higher affinity for the activated (R*) state than the nonactivated (R) state in C6μ cells; for nalbuphine the difference is only 35-fold in favor of the activated state (Emmerson et al., 1996). Thus, DAMGO shifts the equilibrium very much in favor of R*. In contrast, the shift seen with nalbuphine is much smaller. Consequently, the maximal effect of nalbuphine can never be as great as that seen with a maximal concentration of DAMGO, assuming that nalbuphine does not activate as many μ-receptors or the activated receptors cannot access the entire pool of pertussis-toxin sensitive Go/Gi proteins available to DAMGO. Equally well, with lower concentrations of DAMGO the amount of R* will be smaller such that at certain concentrations of DAMGO the number of activated (R*) receptors will be equivalent to the number produced by a certain concentration of nalbuphine. At this point, the extent of activation and the rate of activation will be equivalent for DAMGO and for nalbuphine.
Unfortunately, this same argument cannot be applied to all opioid agonists. For example, the highly efficacious agonist BW373U86 binds equally well to both high- and low-affinity forms of the δ-receptor (Childers et al., 1993) and so would not be expected to select a particular conformation. In addition, both partial agonist and full agonist oripavines bind equally to different affinity states of the μ-opioid receptor (Lee et al., 1999). An alternative explanation would invoke the conformational induction model, whereby binding to an inactive receptor (R) converts this to an active receptor (R*; Burgen, 1981); full and partial agonists produce different proportions of the active conformation(s). However, the level of abundance of active conformation(s) would be the same whether a small amount of full agonist or a large concentration of partial agonist is used. Because both conformational selection and conformational induction models are needed to explain the action of different opioid agonists acting at the same receptor this supports the suggestion that these are extremes of the same mechanism (Kenakin, 1997).
In conclusion, the present results demonstrate that the number of receptors in the same active conformation, or conformations, governs the rate of G protein activation. The data are consistent with the suggestion that receptors can be held in the same active conformation by both full and partial agonists. The difference between opioid partial and full agonists may then be explained by the number of active receptors (of the same conformation) in the presence of a particular drug.
We thank Drs. H. Akil and A. Mansour for providing the rat μ- and δ-receptor clones and Drs. R. R. Neubig and A. Bertalmio for excellent discussion.
This work was supported by National Institutes of Health Grants DA 00254 and DA 04087.
- C6 glioma cell line stably expressing the rat μ-opioid receptor
- C6 glioma cell line stably expressing the rat δ-opioid receptor
- double-distilled water
- Received July 31, 2001.
- Accepted October 5, 2001.
- The American Society for Pharmacology and Experimental Therapeutics